1582
EUGENEI. SNYDER AND JOHN D. ROBERTS
ments verified this prediction. No information as to the relative sign between JABand JAX could be obtained from the complex patterns observed. In principle i t should be possible to obtain the relative sign between JAX and JBXby irradiation between A and B, but this is not feasible experimentally because the chemical shift difference is only of the order of 25 c.P.s.; thus the two decoupled AB regions will overlap in the audio sideband phase detection technique.5 Discussion Although for some time i t was considered that proton-proton coupling constants were all of the same sign, i.e., positive, this notion was overthrown first experimentally by Alexanders and since by other workers1° and theoretically by the work of Karplus3 on the n-electron contribution to the spin coupling constants. I n the present work i t was determined that JAB has a different relative sign from JBXin l-chlorobutadiene-1,2, but the problem of the absolute signs cannot be settled experimentally by the (9) S. Alexander, J . Chcm. Phys., 2S, 358 (1958), 32, 1700 (1961). (10) See for example: J. A. Elvidge and L. M . Jackman, Ptoc. Chcm. Soc., 89 (1959); C. N . Banwell, A. D. Cohen, N. Sheppard and J. J. Turner, i b i d . , 266 (1959); R . W. Fessenden and J . S. Waugh, J . Chem. Phys., SO, 944 (1959); A. D . Cohen and N . Sheppard, Proc. Roy. SOC.(London), 8152, 488 (19BY); F. S. Mortimer, J . Mol. Spec., 8, 335 (1959); P. T. Narasimhan and M. T. Rogers, J . Chcm. Phys., 33, 727 (1960); E. 0. Bishop and R . E. Richards, Mol. Phys., 9 , 114 (1960); C. N.Banwell and N. Sheppard, ibid., 3, 351 (l9GO); A. A. Botbner-By and C. Naar-Colin, J . A m . Chem. SOC.. 83, 231 (1961).
[CONTRIBUTION KO.2741
FROM THE
~701.
8.1
method used. Snyder and Roberts’ have been able to determine that J A X and J S X have the same signs from high resolution analysis, but they could not determine whether the latter two coupling constants have the same sign as J A B . If one uses the theoretical formulations developed by Karplus for determining the u-electron” and a-electron3 contributions to the coupling constants, then an absolute assignment can be made. I t would seem reasonable to take J n x as due primarily to a-electron contribution of the type HCCH which varies from about 0 to +18 C.P.S. depending on the angular disposition of the two C-H bonds. The n-electron contribution for a system H-C= C=C-H (JAB) as calculated by Karplus is about -6.7 c.P.s.~ From the double irradiation experiments reported here we found that JABhas a different sign from J B X which is consistent with Karplus’ theoretical work.3 From our work and that of Snyder and Roberts’ the relative signs of all the coupling constants can be assigned. Then on the basis of Karplus’ theoretical work the signs of JAXand JSXshould be positive and the sign of JAB negative. Acknowledgments.-The authors are indebted to Dr. R. Freeman for information on his use of spin decoupling for the determination of the relative signs of proton-proton coupling constants. Lt’e are grateful to Dr. E. I. Snyder and Professor J. D. Roberts for the sample of 1-chlorobutadiene-l,2. (11) M. Karplus, J . Chcm. Phys., 30, 11 (1959).
GATESA N D CRELLINLABORATORIES OF CHEMISTRY, CALIFORNIA INSTITUTE OF TECHNOLOGY, PASADENA,
Long-Range Nuclear Spin-Spin Coupling.
CALIF.]
Allenic and Polyacetylenic Systems
BY EUGENE I. SNYDER~ AND JOHN D. ROBERTS RECEIVED OCTOBER12, 1961 Spin-spin coupling between protons separated by 5 to 9 chemical bonds has been observed in allenic and polgacetylenic systems. Analysis of the n.m.r. spectrum of l-chloro-1,2-butadieneby various techniques permits assignment of a sign to Relation of these results t o the theory of proton-proton spin-spin coupling is discussed. J13 opposite t o that of J L 2and J,,.
Theoretical studies of the mechanism(s) of nuclear spin-spin coupling constants have been rather successful in elucidating the quantitative aspects of some types of proton-proton spin coupl[n.g. In particular, Karplus and co-workers have utilized the groundwork by Ramsey3 and McConnel14 to devise means which enable correct predictions of both the magnitude and the sign of several types of proton-proton coupling conparticularly for unsaturated organic molecules. Recent reports of considerable long-range couplings in saturated systems6 and of possibly (1) Supported in part by the Office of Naval Research. (2) Postdoctoral Fellow of the Division of Medical Sciences of the
United States Public Health Service. (3) N . F . Ramsey, Phys. Rco., 91, 303 (1953). (4) H . M. McConnell, J . Chem. Phys., 24, 460 (1956). (5) (a) For a summary, see M. Karplus, ibid., 64, 1793 (1960); (b) M. Karplus, ibid., 33, 1842 (1960); ( c ) H . S. Gutowsky, M. Karplus and D . M. Grant, rbid., 31, 1278 (1959): (d) M. Karplus, ibid., 30, G (1959); (e) cf. also H . Conroy, A h . i n Orp. Chcm., 2, 265 (1960). (6) (a) J. Meinwald and A. Lewis, J . A m . Chcm. Soc., 83, 2770
negative 1,2-coupling constants in alkyl derivatives’ may well diminish the degree of faith with which the Karplus approach may be applied to saturated molecules. The utility for unsaturated systems has so far remained unquestioned. An advantage of the Karplus approach to spin coupling for organic chemists is the manner in which i t permits qualitative conceptualization of the nature of proton-proton spin coupling within the framework of valence-bond theory. Thus, compounds for which valence-bond structures with a hydrogen-hydrogen bond may be written in the manner of “second-order hyperconjugation”8 can (1961): (b) D . R. Davis, R . P . Lutz and J. D . Roberts, ibid., 8 S , 246 (1901): ( c ) C. A. Reilly and J. D . Swalen, 1.Chem. Phys., 34, 980 (1961); (d) F . A. L. Anet, Con. J . Chem., 39, 789 (1901); (e) J. V. Hatton and R . E. Richards, Mol. Phys., 3, 253 (1900). (7) (a) F . Kaplan and J. D. Roberts, 1. A m . Chem. Soc., 83, 4607 (1961); (b) C. A. Reilly and J. D . Swalen, J . Chem. Phys , 35, 1522 (1961); ( c ) R. R . Frazer, R. U . Lemieux, and J. D . Stevens, J . A m . Chrm. SOC.,83, 3901 (1961).
SPIN-SPINCOUPLING IN ALLENESAND ACETYLENE~S
May 5, 1962
-H
H-
3600
VI\
vB
.
35100
3400
1583
3300
neb0
IO9 3
997
-‘
355 0 331 8
Zlb.0
en0.o
e3ao
z4b.o
I/
-
Ya
= 237.40
Ye
221.40
vx
69.73
Fig. 1.-Experimental n.m.r. spectrum at 60 mc. for neat l-chloro-1,2-butadiene (top) and spectrum calculated for the parameters indicated (bottom).
be expected to exhibit a t least some degree of spin coupling between the two relevant protons. The predicted magnitude of the coupling constant depends upon the importance of the resonance structures of this type to the over-all resonance hybrid while the sign of the coupling constant is obtained from the algebraic summation of the coupling constants deduced from the individual structures. The coupling constants associated with particular structural features have been in part evaluated from electron-proton hyperfine interactions in paramagnetic resonance studies. The purposes of the present work were to determine the magnitude and sign of coupling constants in allenic derivatives and to investigate some unsaturated systems for which significant “super long-range” coupling constants might reasonably be expected.
Fig. 2.-Experimental n.m.r. spectrum at 40 mc. for neat l-chloro-1,2-butadiene (top) and spectrum calculated for the parameters indicated (bottom).
*-H
H-1
B
Results The coupling constants found for the various compounds which were investigated are summarized in Table I. Coupling constants between equivalent protons and between the accidentally equivalent acetylenic and methyl protons in 1,3-pentadiyne were obtained from the 13C-satellitespectra. All of the coupling constants except those for 1chloro-l,2-butadiene could be obtained by simple first-order interpretation of the relevant spectra. The coupling and chemical shift parameters for l-chloro-1,2-butadiene, an ABXs system, were obtained in the following manner. The relative was determined from the behavsign of JAX and JBX ior of the XSresonance signals a t 40 and 60 mc., for the theory of ABXI systems demands that those lines whose separation is 1 JAX J B XI be field invariant.’O From Figs. 1 and 2 i t is apparent that ~ J A Xf J B X=~9.6 C.P.S. Since the data from other allenes and alkenes possessing the CH3CH= grouping indicate that JAX and JBX should be about 2-3 and 7 c.P.s., respectively, the value of
+
(8) G. W. Wheland, “Resonance in Organic Chemistry,” John Wiley and Sons, Inc., New York, N. Y . , 1955, pp. 160-152. (9) A. D. Cohen, N. Sheppard and J. J. Turner, Proc. Chcm. Soc., 118 (1958). (10) V. J. Kowalewski and D. G. de Kowalewski, J . Chcm. Phys., BS, 1794 (1960).
Fig. 3.-Methyl group p ttern observed in doubl resonance experiments with sideband of 258 C.P.S. (left) and 245 C.P.S. (right).
+
~ J A X Jbxl implies that both coupling constants are of the same sign. Further support for this assignment is given by the outward movement of the inner Xa lines with increasing magnetic field, behavior which is characteristic of an ABX3 system when ~ J A X J B X> ~ 1 . 7 ~~ Jsx\.’O T h a t the sign of JM, is opposite to that of JAx was unequivocally established from double-resonance experiments of the type described by Freeman,” performed for us by Drs. Dan Elleman and S. L. hianatt of the Jet Propulsion Laboratory.
+
(11) R . Freeman, Mol. Phys., 8, 435 (1960); R. Freeman end D. € Whitlen, I. ibid., 4, 321 (1961); R . Freeman, ibid., 4, 385 (1961). See also J. P. Maher and D. F. Evans, Proc. Chcm. Soc., 208 (1961).
EUGENEI. SNYDER AND
1584
JOHN
D. ROBERTS
Vol. 84
TABLE I COUPLING COSSTANTS IS SOME UNSATURATED SYSTEMS Compound
Solvent
( CH3)2CzC=CH2 ( CH3)zC=C=CHCl
Neat Seat Xeat
CHaCH=C=CHCl Ss B A
Jlaos, C.P.S.
JH-H, C.P.S.
J I= ~ 3.03 i 0.06 J14 = 2 14 i .OS J A B = J:, F6.8
166" Z!Z
0.1
ClCH:C=CC=CCH2CI Seat Jlb = 1 . 0 zk .2 158 CH~CICCSCCECCH~OH CHBOH JlS = 0 . 4 For CHI grouping. M. Karplus, D. H . Anderson, T. C. Farrar and H. S. Gutowsky, J . Chem. Phys., 27, 597 (1967). J. N. Shoolery, L. F. Johnson and W. A . Anderson, J . Mol. Spectros., 5 , 110 (1960); E. B. Ti'hippIe, J . H. Goldstein and Le. E. Stewart, J . Am. Chem. Sac., 81, 4761 (1959); W. R. l'aughan and R . C. Taylor, J . Chem. Phys., 31, 1425 (1959). S.Muller and D . E . Pritchard, i b i d . , 31, T68 (1951). For CHa grouping.
Final values of the n.m.r. parameters were obtained by an iterative procedure, based on an exact solution of the high resolution n.m.r. Hamiltonian, programmed for an IRA1 7090 computer. Calculated and experimental spectra for both 40 and 60 mc. are shown in Figs. 1 and 2. The extreme ease of polymerization of butatrithwarted our attempts to ene" even a t -28' record the 13C-satellitesof its proton spectrum. This was unfortunate since comparison of the experimental and predictedjb value would have been particularly interesting.
Discussion
A necessary corollary of this would be that a significant hyperfine interaction is associated with the structural grouping CHa
\
.
c=c-.
/
Consideration of the valence bond structures of dialkyl-substituted polyacetylenes of type I1 sug>HC-(C=C),-CH< I1
-H H >C=(C=).=C< IIa
gests that these systems should exhibit long-range couplings. Similarly, the simple and monoalkylsubstituted polyacetylenes (111 and IV) groupings
The spectra of the allenic compounds examined H clearly demonstrate the existence of a 1,4-coupling CZEC),-C=CH f+ : C=C=(C=C),=C=C : of magnitude 2-3 C.P.S. Assuming HC--C-( constant (J14) I11 IIIa that J I 2is positive, as theory p r e d i ~ t s , ~ then ~.~,'~ H----.--H JI4 is positive and the coupling constant between > CH-( C=C),-C=CH +-+ > C=( C=C),2=-C=C : the allenic protons ( J I B )is negative. Our results IY 11:a thus provide experimental verification of Karplus' prediction5b of a negative sign for this latter type might also exhibit substantial long-range coupling. (J1:Jof coupling constant. No theoretical con- The first members of series I1 ( n = 1) and I\' ( n = siderations regarding either the sign or magnitude 0) are known to have their terminal protons spinof J14in allenic systems are available. If one coupled (see Table I). \Ve have prepared several assumes that for compounds of type I that Ia is an representatives of series I1 ( n = " 3 ) and have important valence-bond structure from the stand- demonstrated coupling across T and 9 bonds, repoint of proton-proton coupling, then J14 is pre- spectively ( c j . Table I ) . The terminal protons are also coupled in those representatives of series H H I11 ( n = 0) and IV ( n = 2) examined by us. H Comparison of the J-values for acetylenes is CH,-C=C=CHc-* CH,=C-C=C #CH2-C=C-C: I I I I I illuminating in several respects. Firstly, were X Y X Y X Y a direct-through-space mechanism the important I Ia Ib one for proton-proton coupling, I J , would be indicted to be positive.5b However, the magnitude versely proportional to the third power of the of JI4so calculated, 3.9 c.P.s., is rather larger than distance between protons and would decrease with the experimental value (3.0 c.P.s.). This discrep- n much faster than it actually does. The inapancy suggests that perhaps structure Ib, with a plicability of direct mechanisms to pvoton-proton predicted negative J14, is sufficiently important spin coupling has been frequently cited on theo(in the spin-spin coupling sense) to reduce sub- retical base^,^^^^^^ and the present series of cornstantially the coupling below the calculated value. pounds provides a (probably unnecessary) experimental confirmation of the unimportance of such ( I ? ) W. .\I Schubert, . T. H. Liddicoet a n d 'Ar. A. Lanka, J . A m Chenz SOL.,76, 1929 (1954). mechanisms. (13) S o t e , however. t h a t recent work'x strongly suggests t h a t Hoffman and Gronowitz have predicted14 that JABis indeed of opposite sign t o JAXand J B X in diethyl sulfite Consequently. either 1,l- or 1,2-couplinga seem tive in vuturated cumpounds.
to
be required
to
be nega-
(14) (a) R . -4. Hoffman, J . Mol. P h y s . , 1, 326 (1958); (b) R H ~ 5 m a nand S. Crorrowitz, Acta Chrm. Scand., 13, 1477 (1959).
A.
May 5, 1962
SPIN-SPINCOUPLING IN XLLENESAND ACETYLENES
substitution of a methyl group for an olefinic proton should give essentially the same coupling between some other proton and the methyl group as between the other proton and the olefinic proton, provided that coupling is transmitted by a T electron system. The rationale of this suggestion is that the hyperfine interaction constant for methyl and ethyl radical fragments are nearly identical in magnitude although opposite in sign.1h815The same approach does not seem to apply to the acetylenic series, where substitution of a hydrogen by a methyl group in diacetylene causes considerable decrease in I J1, perhaps because of the lack of a similar relation between the hyperfine interaction constants of higher radical fragments, e:g., HC=C-C and CH3CrC-C. It is possible, if the current theory5bof spin coupling in T-electron systems is valid, that the magnitude of appreciable hyperfine interactions in radicals of this type might be more easily evaluated from n.m.r. than e.p.r. spectra. One point may be of practical importance to the organic chemist who uses nuclear magnetic resonance spectroscopy as a tool for analysis of organic structures. Xlthough the concept of conjugation is highly useful for ultraviolet and infrared spectroscopy, uninhibited extension of the concept to qualitative considerations of spin coupling constants may be misleading. By way of illustration, J14 in the conjugated system, diacetylene, is less than that in butyne-2. Furthermore, J14 in l-chloro-1,2-butadiene, where each double bond behaves as an independent unit, is easily observable, and larger than in some conjugated systems,l6whereas J14 in 3-chloro-1-butyne is not observable. The mere existence or absence of a conjugated system is not enough and one must consider first of all those valence-bond structures to which can be ascribed the origin of spincoupling constants. Caution must be exercised here also, because contributions from various structures can both reinforce or cancel one another. Since hyperconjugstion appears to be almost inextricably bound to the theoretical development of the phenomenon of proton-proton spin-spin coupling,l4 it might be worthwhile to mention that ground-state hyperconjugation can be regarded its highly important for spin-spin coupling even though it is almost certainly relatively unimportant so far as determining the energy of the ground state is concerned.'' It remains to be seen how well spin-coupling constants, as a measure of hyperconjugation (or some similar electronic effect), can be used for correlation of changes in rate or equilibrium constants due to substituent groups. Experimental Suclear magnetic resonance spectra were recorded with a Yarian Associates high resolution n.m.r. spectrometer operated a t 60 mc., unless otherwise noted, and equipped with a Super Stabilizer. Line positions relative to a suit(15) For a summary and references, see R. A. Hoffman and s. Gronowitz, A r k i s Kemi, 16,471 (1960). (16) E. I. Snyder, unpublished work. (17) The ground s t a t e energy of methane is supposedly changed by only 1.3 kcal by inclusion of those structures responsible for spin bouplinp (ref, ad).
1585
able internal standard were measured by either the sideband superposition technique or by interpolation between sidebands placed laterally to the resonance signals. 3-Methyl-l,2-butadiene was prepared by lithium aluminum hydride reduction of 3-chlor0-3-methyl-l-butyne'~ and purified by distillation through a Nestor-type spinningband column followed by redistillation of the fractions of >9470 purity ( n 2 0 ~1.4116-1.4190) through a center-tube fractionating column. A fraction, b.p. 39-40', nzoD 1.4188 (lit.'* b.p. 40.0-40.2O, na5, 1.4148), of 99+% purity (by V.P.C. analysis on a diisodecyl phthalate column) was used for the spectral work. l-Chloro-3-methyl-l,2-butadiene .-Omission of copper bronze from the published procedure19 for preparation of this compound resulted in no deleterious effects, and 57% ~ of crude product of b.p. 40-50" (105-92 mm.), n Z 51.46621.4741, was obtained. Careful fractionation through a Nestor-type spinning-band column of a high-boiling cut, n 2 5 ~1.4741, gave a sample of b.p. 51.8-52.0" (110 mm.), n 2 5 ~1.4744 [lit.'9 b.p. 60-63' (175 mm.), nZ5D 1.47401, 99.570 pure by v.P.c., which was used for the spectral work. l-Chloro-l,2-butadiene.-Treatment of 3-butyn-2-01 with thionyl chloride in diethyl Carbitol afforded a mixture of 1chloro-1,2-butadiene and 3-chloro-l-b~tyne.~~ The allene was separated by preparative v.p.c.2' and was 99.7% pure by v.p .c. 3-Chloro-1-butyne was the major component ~ of the reaction mixture and when separated had n 2 51.4214.
Anal. Calcd. for C4HaCl: C, 54.26; H, 5.69; C1, 40.05. Found: C, 53.93; H, 5.86; C1,39.64. ,4n attempt to prepare l-bromo-1,2-butadiene by an s N 2 ' displacement with lithium bromide on the tosylate of 3-butyn-2-01 afforded only the acetylenic bromide. Diacetylene generated from 1,4-dichlor0-2-butyne-2~~ was purified by sublimation a t -30" followed by a bulb - tobulb distillation. A sample was sealed into an n.m.r. tube a t 1 mm. and stored a t -78". Spectra were recorded a t the ambient temperature (-28"). 1,3-Pentadiyne.-An explosion occurred in an attempt to prepare this compound by the method of Armitage and c o - w o r k e r ~ . ~Accordingly, ~ the isolation procedure was modified slightly. The material obtained by concentration of the butane extract of the reaction mixture was filtered and distilled a t water-pump pressure. The distillate was collected a t - 78" and then rapidly distilled on a steam-bath in an atmosphere of prepurified nitrogen to afford 1,31.4762, 1.4770 (lit. 7 2 ' 5 ~ pentadiyne in two fractions, 1.4790,241.471723). Extensive decomposition was found to accompany distillation. -4 portion of the product was distilled under reduced pressure into an n.m.r. tube containing benzene and stored in nacuo in the dark a t -78' until its spectrum was recorded a t room temperature. 2,4-Hexadiyne .-The reported procedure afforded material of m.p. 65.0-65.8' (sealed capillary, lit.22m.p. 67"). Its n.m.r. spectrum was determined for a 5Oy0 solution in carbon tetrachloride. l,6-Dichloro-2,4-hexadiyne.-Treatment of 2,4-hexadiyne-1 ,6-diolZ2in pyridine with thionyl chloride afforded ~ yellowish dichloride of b.p. 55-58' (0.5 mm.), n Z o1.5749 1.57701. The product [lit.25 b.p. 61" (0.5 mm.), was stored a t 0' prior to spectroscopic investigation. 2,4,6-0ctatriyn-l-o1 was prepared from bromopropargyl alcohol and 1,3-pentadiyne as described by Chodkiewicz.26 Two recrystallizations in the dark from carbon tetrachloride gave material of m.p. 92-93' (lit.27m.p. 93n). Because of (18) W.J. Bailey and C . R . Pfeifer, J. Org. Chem., 20, 95 (1955). (19) G.F.Hennion, J. J. Sheehan and D. E. Maloney, J. A m . Chem. SOL.,72, 3542 (1950). (20) T. L. Jacobs, W. L. Petty and E . G. Teach, rbid., 82, 4094 (1960). (21) We wish t o thank hIr. Kent Smitheman of American Potash Co. for performing this separation. (22) J. B. Armitage, E. R. H. Jones and hr. C. Whiting, J. Chem. Soc., 44 (1951). (23) J. B. Armitage, E. R. H. Jones and h'f. C. Whiting, ibid., 1993 (1552). (24) H.Schlubach and V. Wolf, Ann., 668, 141 (1950). (25) C. L. Cook, E. R . H . Jones and h l . C. Whiting, J. C k r m . S O L . ,2883 (1952). (26) W. Chcdkiewtcz, Ann. shim., (131 a , 852 (1967).
1586
KENNETH D. KOPPLE
the extreme photosensitivity and the line-broadening effect of the photolysis products on the n.m.r. spectrum, a sample was sublimed in the dark a t 0.5 mm. immediately prior
[CONTRIBUTION FROM
THE
Yol. s4
to spectral study. The resulting material was pure white; it was dissolved in methanol, stored a t Oo, and protected fromlight.
DEPARTMENT OF CHEMISTRY, UNIVERSITY OF CHICAGO, CHICAGO 37, ILL.]
Reduction of Unsaturated Compounds by Ammoniacal Chromium (11) BY KENNETH D. KOPPLE RECEIVED OCTOBER 25, 1961 Chromium( 11) in concentrated aqueous ammonia has been shown to convert benzaldehyde and acetophenone to carbinols and to saturate the ethylenic double bond of cinnamic acid and mesityl oxide. Differences between this reagent and aquo chromium( 11) are discussed.
Chromium(II), while stable in oxygen-free aqueous acid, has been observed to evolve hydrogen from alkaline solutions containing excess oligopolyethylenimines. This observation suggested that i t might be worthwhile to establish how other reducing properties of chromium(I1) ammines differ from those of the aquo ion. Almost 100 years ago Berthelot reported that ammoniacal chromium(I1) chloride converted acetylene t o ethylene,2 but with the exception of some experiments with chromium(I1) hydroxide3 subsequent studies of the action of divalent chromium on unsaturation have been limited to reaction in acidic solution, Solutions of chromium(I1) ion in concentrated aqueous ammonia may be stored under nitrogen for weeks without apparent decomposition, but they react almost instantaneously with a number of a,Bunsaturated carbonyl compounds that are unaffected by acidic solutions of chromium(I1) salts. In their survey of the reactivity of the latter solutions toward carbonyl ~ o m p o u n d s ,Conant ~ and Cutter found that while acetylenic derivatives, quinones and maleic and fumaric acids underwent two-electron reduction, where reaction occurred with other a,/?-unsaturated carbonyl compounds one-electron products were obtained. For example, benzaldehyde and other aromatic aldehydes were converted to hydrobenzoins. (Acetophenone and saturated aliphatic aldehydes were not attacked.) Acrolein was reduced to 3,4-dihydroxy-l,5hexadiene and benzalacetone to 4,5-diphenyl-2,?-octadione. (Mesityl oxide and cinnamic acid resisted attack.) In contrast, ammoniacal chromium(I1) sulfate affords, upon extraction of the reaction mixture, two-electron products from benzaldehyde (benzylamine, isolated in 31YG yield as the Nphenylurea), acetophenone (a-phenylethyl alcohol, isolated in 71% yield as the N-phenylcarbamate), mesityl oxide (methyl isobutyl ketone, isolated in 33YGyield as the 2,4.-dinitrophenylhydrazone) and cinnamic acid (hydrocinnamic acid, isolated in 45y0 yield). A related example, the reduction of cinnamic to hydrocinnamic acid by chromium(I1) hydroxide, has been reported. Other substances, (1) K. D. Kopple, G. F. Svatos and H. Tsube. Nature, 189, 393 (190 1). (2) hf. Berthelot, Ann. chim., 141 9, 401 (1868). (3) W. Traube snd W.Passarge, Bcr., 49, 1892 (1916). (4) J. B. Conant and H. Cutter, J . A m . Chem Soc.. 48, 1016 (1926).
not affected by the acidic solutions, that rapidly oxidize the amnioniacal reagent include acrylonitrile, acrylic acid and acrylamide. Neither reagent affects aliphatic ketones or cinnamyl alcohol. When reaction occurs with the ammoniacal reagent, i t is rapid and, where excess oxident is used, chromium(I1) is consumed completely. Low yields of isolated product occur because of the formation of stable, unextractable chromium(II1) complexes of the organic products. (A quantity of organic-soluble complex of hydrocinnamic acid was obtained; on decomposition i t afforded an additional 17y0of the acid.) Conant and Cutter apparently had the same difficulty with acidic chromium(I1) ; they obtained cleaner organic products when vanadium(I1) was used.
Experimental Materials and Methods.-For reduction experiments on a preparative scale a 0.8 M solution of chromium(I1) sulfate in 1.2 N sulfuric acid, obtained by zinc reduction of a chromium( 111) sulfate solution, was employed. This was stored in a serum-capped flask, under nitrogen, and transferred by means of a hypodermic syringe to reaction vessels which were likewise capped. The contents of these flasks had been freed of air by flushing with nitrogen before addition of the reductant. For other experiments a stable solution of chromium( 11) chloride in concentrated aqueous ammonia was prepared by solution of the anhydrous salt in concentrated aqueous ammonia. Screening experiments were carried out in serum-capped test-tubes using this reagent. The occurrence of reaction was assumed to be indicated by a color change from deep blue to orange or violet. Spectra were measured by means of a Cary model 14 spectrophotometer. Ammoniacal chromium(I1) chloride was added to a serum-capped test-tube containing a n excess of the pure oxidant. After reaction had occurred, the tube was immersed in a Dry Ice-acetone-bath until the contents were frozen and then mixed with sufficient 6 N HCI to give 1-2 M excess hydrogen ion. When an oxygenfree atmosphere was required the thawed contents of the reaction tubes were transferred by syringe to serum-capped, nitrogen-flushed cuvettes for measurement. Substances Screened.-The following substances, used in excess, reacted within seconds with ammoniacal chromium(I1) chloride or sulfate: cinnamic acid, maleic acid, acrylic acid, acrylarnide, acrylonitrile, benzaldehyde, acetophenone, mesityl oxide, p-benzoquinone. Unreactive on 24-hr. storage with the reagent were cinnamyl alcohol, hexamethylenetetramine, methyl ethyl ketone, diethyl ketone, phenetole. In one experiment 0.5 ml. of freshly distilled acrylonitrile was shaken, under nitrogen, with 4 ml. of 0.1 M chromium(11) chloride in concentrated ammonia. The clear orange solution which resulted gradually reddened, and after 24 hr. was reddish-violet and still free of precipitate. It was acidified and dialyzed against water. Material remaining in the
